While few biologists or philosophers have taken these pronouncements particularly seriously, some commentators have begun to see in the development of big data methods in the biological sciences the possibility that future biology will be able to do entirely without theory, simply drawing statistically robust generalizations from vast corpora of data and operating without understanding of underlying mechanisms or causation. To explore this claim in more detail, I’ll unpack the idea of “data integration” that lies behind the construction of many of these big-data explanations, arguing first that data integration in this sense is not a species of the kind of inter-domain integration that I’ve developed here, and then that, while we might construct a plausible version of the argument in favor of theory-free biology, criticisms of that notion that have been advanced by philosophers clearly show us that it lacks integrative promise.